Vision-based Non-contact Detection Method Of Heart Rate And Respiratory Rate

With the development of the national economy and the improvement of people’s living standards,the awareness of health care for the whole nation has been continuously enhanced,and people are increasingly demanding to understand their own physiological conditions at any time in daily life.As an important parameter reflecting human physiology,heart rate and respiration rate can provide important information about the health of the human body.Therefore,it is important to achieve daily monitoring of heart rate and respiratory rate.However,most of the technologies for detecting heart rate and breathing on the market are contact type,and it is necessary to connect the test subject with the measuring device,which has the disadvantages of restraint,complicated operation and high price,which is not conducive to daily monitoring.Therefore,the use of non-contact sensing methods,seeking for long-term monitoring,can be used for disease prevention,easy to operate,low-cost monitoring methods,has become the mainstream research direction at home and abroad.Based on the image photoplethysmography and Euler’s video amplification principle,this paper takes the non-contact measurement of heart rate and respiration rate as the research goal,and studies the extraction method of heart rate and respiration rate based on common camera,which has important theoretical significance and Practical application value.Based on the Euler amplification algorithm,this paper uses the ordinary camera to capture the human respiratory video,and uses the Euler algorithm to amplify the chest and abdomen movement displacement during breathing,and considers the effect of the location accuracy of the chest and abdomen region on the detection accuracy of the respiratory rate.A method for extracting the respiratory region based on the optical flow signal is proposed.The optical flow algorithm is used to convert the chest and abdomen motion into optical flow information and encode it,which is displayed as a color image form,and the pixel average brightness sequence of the chest and abdomen breathing region is extracted.Thereby,respiratory waveform information is obtained,and the respiratory rate is obtained by peak detection.In this paper,based on the image photoelectric volume pulse pulsing method,the pulse wave signal is extracted from the captured face video by using the independent component analysis method,and the heart rate is obtained by analyzing the power spectrum of the pulse wave.However,the tiny motion of the face in the video has a great interference with the pulse wave signal,and the accuracy of the pulse wave signal obtained from different regions of the face is also very different.In order to suppress the influence of motion and face area on heart rate measurement accuracy,this paper uses DRMF algorithm to detect the position of key points of the face,and divides the face into four positions according to the key points: forehead,cheek,mouth and whole face area.The selected area is tracked in real time,and the pulse wave signals acquired in the four regions are analyzed.Finally,the frequency domain signal of the pulse wave signal is obtained by the fast Fourier transform to obtain the heart rate.In the final combination experiment,the heart rate results extracted in this paper are compared with the Euler amplification algorithm and the standard heart rate results of the polysomnography;the respiratory rate extracted in this paper is compared with the standard results measured by polysomnography.The experimental results were analyzed by using the average absolute error Me,the root mean square error RMSE and the Pearson correlation coefficient r.The results show that the average error of heart rate detection and respiratory rate detection is 1.16 beats/min and 0.54 beats/min.High accuracy.